Proc. Nati. Acad. Sci. USAVol. 88, pp. 10647-10651, December 1991Evolution

Evidence for common ancestry of a chestnut blight hypovirulence-associated double-stranded RNA and a group of positive-strandRNA plant virusesEUGENE V. KOONIN*t, GIL H. CHOIt, DONALD L. Nussf, RONI SHAPIRAf, AND JAMES C. CARRINGTON*§*Department of Biology, Texas A&M University, College Station, TX 77843; and tDepartment of Molecular Oncology and Virology, Roche Institute ofMolecular Biology, Roche Research Center, Nutley, NJ 07110

Communicated by Myron K. Brakke, August 1, 1991 (received for review June 1, 1991)

ABSTRACT Computer-assisted analysis of the putativepolypeptide products encoded by the two open reading framespresent in a large virus-like double-stranded RNA, L-dsRNA,associated with hypovirulence of the chestnut blight fungus,Cryphonectria parasitica, revealed five distinct domains withsignificant sequence similarity to previously described con-served domains within plant potyvirus-encoded polyproteins.These included the putative RNA-dependent RNA polymerase,RNA helicase, two papain-like cysteine proteases related to thepotyvirus helper-component protease, and a cysteine-rich do-main of unknown function similar to the N-terminal portion ofthe potyvirus helper-component protein. Phylogenetic treesderived from the alignment of the polymerase domains ofL-dsRNA, a subset of positive-stranded RNA viruses, anddouble-stranded RNA viruses, using three independent algo-rithms, suggested that the hypovirulence-associated dsRNAand potyvirus genomes share a common ancestry. However,comparison of the organization of the conserved domainswithin the encoded polyproteins of the respective viruses indi-cated that the proposed subsequent evolution involved exten-sive genome rearrangement.

The phenomenon of transmissible hypovirulence representsa natural form of biological control in which the virulence ofCryphonectria parasitica, the chestnut blight fungus, is mod-ulated by the presence of virus-like genetic elements com-posed of double-stranded RNA (dsRNA) (reviewed in refs. 1to 4). Efforts to understand the molecular mechanisms re-sponsible for transmissible hypovirulence have provided anemerging view of the structural and functional properties ofhypovirulence-associated dsRNA genetic elements (re-viewed in ref. 5). The largest dsRNA present in C. parasiticastrain Ep713, L-dsRNA (12,712 base pairs), was shownrecently to contain two contiguous coding domains, desig-nated open reading frame (ORF) A and ORF B, consisting of622 and 3165 codons, respectively (6). Both ORFs encodepolyproteins that undergo autocatalytic processing during orimmediately after translation (6, 7). On the basis of thesimilarity of the L-dsRNA genetic organization and expres-sion strategy to those of several viral genomes, it wassuggested that L-dsRNA should be considered the equivalentof a viral genome or replicative form and the descriptive termhypovirulence-associated virus (HAV) was proposed (6).

Similarities between one of the HAV-encoded proteases,p29, and the potyvirus-encoded protease, HC-Pro, werenoted previously (7). In addition, computer-assisted analysisof the C-terminal portion of the ORF B-encoded polyproteinrevealed a domain that was clearly related to the RNAhelicase of potyviruses (6). We now extend these observa-tions by demonstrating that three additional domains are

conserved between the gene products of HAV and those ofpotyviruses, including the putative RNA polymerase. Adetailed analysis of the sequence and organization of theconserved domains within the HAV-encoded and potyvirus-encoded polyproteins suggests that Ep713 HAV dsRNA andthe positive-stranded RNA genomes of the potyviruses sharea common ancestry and that their evolution included exten-sive gene shuffling.

MATERIALS AND METHODSPolyproteins encoded by HAV RNA, barley yellow mosaicvirus (BaYMV) RNA 1, BaYMV RNA 2, and pea seed-bornemosaic virus (PSBMV) RNA were from refs. 6, 8, 9, and 10,respectively. All other sequences were taken from theSwissprot data base (release 16). Methods used for multiplesequence alignment, cluster dendrogram formation, and par-simony tree analysis were described in detail by Dolja et al.(11) and in references therein.

RESULTS AND DISCUSSIONHypovirulence-Associated dsRNA Encodes a Putative RNA-

Dependent RNA Polymerase Related to the Polymerases ofPositive-Strand RNA-Containing Potyviruses and BaYMV.Considering the similarity between the putative helicase ofHAV and the potyvirus-encoded helicase (6), we first com-pared the sequences of HAV gene products with those ofpotyviruses and BaYMV. The extensive sequence similari-ties, despite differences in genome organization, between theproteins encoded by potyviruses and BaYMV have beendescribed (8, 9, 12). Inspection of the local similarity plotsrevealed one of the conserved motifs of the RNA polymer-ases of the positive-strand RNA viruses (designated motif IVin Fig. 1) as the region of highest similarity over the entirelengths of the HAV ORF B polyprotein and the polyproteinsencoded by potyviruses and BaYMV, while motif V scored

Abbreviations: dsRNA, double-stranded RNA; ORF, open readingframe; HAV, chestnut blight fungus hypovirulence-associated virus;HC, helper component. Potyviruses: PPV, plum pox virus; PSBMV,pea seed-borne mosaic virus; PVY, potato virus Y; TEV, tobaccoetch virus; TVMV, tobacco vein mottling virus. Potyvirus-like virus:BaYMV, barley yellow mosaic virus. dsRNA viruses: ScV, Saccha-romyces cerevisiae virus; BTV, bluetongue virus; ROTA, humanrotavirus SA11; REO, reovirus type 3; IBDV, infectious bursadisease virus. Togaviruses: SFV, Semliki forest virus; SNBV, Sind-bis virus; RRV, Ross river virus; ONNV, O'Nyong-Nyong virus;VEEV, Venezuelan equine encephalitis virus. Picornaviruses: PV1,poliovirus type 1; EMCV, encephalomyocarditis virus; FMDV,foot-and-mouth disease virus type A10; HA, hepatitis A virus strainLA. Comovirus: CPMV, cowpea mosaic virus. Nepovirus: GCMV,Hungarian grapevine mosaic virus. Nodavirus: BBV, black beetlevirus.tPermanent address: Institute of Microbiology, Academy of Sci-ences, 117811 Moscow, Russia.§To whom reprint requests should be addressed.

10647

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Apr

il 29

, 202

0

10648 Evolution: Koonin et al. Proc. Natl. Acad. Sci. USA 88 (1991)

la

MGKKKRDILGMRRTDVQGKRRGALEKLTKRDLKEKGKRRFLGKGKHPY

I IKGKKKEALSGKKKQYFQGKKRDLC

FYKSRKAL

PV1 -TGSAVGEMCV -LGSAIGFMDV -IGSAVGHA -TGVAIGCpMV -LSCQVGGCMV -LPCQVG

TEV -APWTVGTVMV -CPWSVGBaYMV -APHTVG

JAV GSGMPL-6.

ScV THKFPVG

LNK-QTR---------DTKEMQ---KLLDT-YGI--NLPLVTVDW-ESATLI----PFAAERLR---KMNEG-DFS--EWYQTIDF-ENGTVG----P-EVEAAL---KLMEK-REY--KFVCQTIWL-DENGLLLGVHPRLAQRILFNTVMMENCSDL--DWFTTVQG-DDCVVSL---IPGTTVAKAYEELEASAHRFVPALVGIEFEECEDGSLKL---KEGSQAAELYENMAQFAKEEVPELVVIE

I I ISELTLDEQEA----MLKASCLR---LYTGK-LG-----IWNGEDLSDDAVAN----LVQKSCLR---LFKNK-LG-----VWNGAHLSDDEVLH----LAEVCRQQ---FLEGKSTG-----VWNG

.A . . 66.Y.KQAGVMDVIR----KNALECIS.--TGKYPT----QFYHA

m

CDP-DLFWSKICDP-DVHWTAFCNP-DVDWQRFIDP-DRQWDELINPYSMEWSRLTNPYSREWGHLMQ I

MTKFYGGWNELMTKFYGGWNELINKFGRGWEKL

* .66.6#SQSMARIWDEL

DQAEAAKVHKR

GIDL--DHLK----NFEF--DDVK----GVEL--DTYT----SPVFFCQALK----APELMVQSFDKLIGTPE--KNSFGINVC

N----KEEIV----DINSQEDVCK----ELALLNERFI----

.#KDFFCSNR.---

ScV DVQD---------

PVLME---EKL-FGVAMQG-FERV-YGTHFAQ-YRNV-WFKTMIR-FGDVGLAARMKE-KGNDVLLNRLRRVKTNEAI

I IMEAL-P-SGWVYCLGKL-P-DGWVYCHDKLNR-PGWLHG666§. 6HD-LRKREGGQFI

VNMMKD--GASSF

VI

MIAYGDDVIASVLSYGDDLLVAMISYGDDIVVAILCYGDDVLIVLVTYGDDNLISLLVYGDDNLIS

ItYYVNGDDLLIAFFANGDDLIIAFVCNGDDNKFA

LYNTSDDTVWW

SVHNGDDVMIS

IV

AFDYTGYDASLDVDYSNFDSTHDVDYSAFDTNHDLDFSAFDASLCCDYSSFDGLLNCDYSGFDGLL

II I IIIDADGSQFDSSLDADGSQFDSSLSGDGSRFDSSI.#.. .#

IADATAYDSNC

CFDYDDFNSQH

SPAWFEA 16aaSVAMFRL 19aaCSDAMNI l9aaSPFMIRE 19aaSKQVMDV 22aaTPQLVEM 25aa

TPFLINA 22aaSPYLINA 22aaDPLFFDV 20aa

KPALFHG 153aa

SIASMYT 29aa

I

YVKDELRSKTKFLKDELRPIEKFLKDEIRPMEKCPKDELRPLEKCPKDEKLPMRKCPKDELLPARK

I I ISLKAELRPIEKSLKAELRPFEKSLKAELRTIEKt .

FAKSQAVPGQP

VNKMPKHKISR

VEQGK-S--VQAAK-T--VRAGK-T--VLESK-T--VFDKPKT--IKVGP-C--I IVENNK-T--LIENK-T--VEAEK-T--* 6.-LLAPRMKDL

MIASP-PEV

IDYLNHSHH-LYKNKTYCVLESLAISTH-AFEEKRFLILKTLVNTEH-AYENKRITVINTIIYSKH-LLYNCCYHVLMACCSRLA-ICKNTVWRVIMALCGRYALVGTQVVYKV

II I IYTEIVYTPILTPDGTIIKKYTEIVYTPISTPDGTIVKKYDEILNTTICLANGMVIKK

IKRIVFANR-EVISNVHHK

SVRHMWVLD-PDTKEWYRL

YPHEVD--A--SLLA-------QSGKDYGLTMTPA-DKSA-TFE--TVTWENTNYQLD--F--DKVR-------ASLAKTGYKITPA-NTTS-TFP-LNSTLEDSDYDLD--F--EALK-------PHFKSLGQTITPA-DKSDKGFV-LGHSITDFSRDVQ--I--DNLDLIGQKIVDEFKKLGMTATSA-DKNVPQL----KPVSEVNAWTPYFDGKKLK-------QSLAQGGVTITDGKDKTSLELP--FRRLEEVSPAVASWFTGEAIR-------VTLAEKRIKITDGSDKDAPTIE--AKPFSE

IHPDKA-----ERLS--------RFKESFGELGLKYEFDCTT-----RDKTQISPELE-----HVLD--------GFQQHFSDLGLNYDFSSRT-----RDKKEISPQFD-----EEFGH-------DFSPELVELGLTYEFDDIT-----SDICE

* 6 .#.6 66§SKDLLS----SAEVD--------RFKQAAADFGILLEIGST------KKITE

LNRVSTAVRIMDAMHRINAR---AQPAKCNLFSIS-EFLRVEHGMSGGDGLG

H

RLIEASSLN----DSVAMRMAFGNLYAAFRIVDVPPFE----HCILGRQLLGKFASKFRIVDVLPVE----HILYTRMMIGRFCAQMRAIDAKPLD----YSILCRMYWGPAISYFRCFTILPME----YNLVVRRKFLNF-VRFRLFEIMPLH----YNLLLRVKTCAF-TAFI I I IRTFTAAPID----TLLAGKVCVDDFNNQFRTFTAAPIE----TLLGGKVCVDDFNNHFRVFTASPIT----SLFAMKFYVDDFNKKF* .. . . *6.RTVVSEDLS----AYMVDQIFQIEANKRI# 6RAWTSTKYEWGKQRAIYGTDLRSTLITNF

V

HK--NPGVIQT--QPGLEHS--NNGPQHL--NPGFHIM--ANRHRLQ--HNRHR

I-Y--DLNIK-Y--SKHIQ-Y--ATNLK

TW--ETYGA

AMFRCEDVL

:* *:KGGMPSGCSGTSIFNSMIN-NLIIRTLLLKTYTG..LPSGCAATSMLNTIMN-NIIIRAGLYLTYEGGMPSDCSATGIINTILN-NIYVLYALRRHYCGSMPSGSPCTALLNSIIN-NVNLYYVSKIIFECGIPSGFPMTVIVNSIFN-EILIRYHYKKLMNCGLPSGFALTVVMNSIFNEEILIRYAYKTL-

I III I III IHKGNNSGQPSTVVDNTLMV-IIAMLYT-CEKCFKGNNSGQPSTVVDNTLMV-VLAMYYA-LSKLNVGTXSGQPSTVVDNTLVL-MTAFLYAYIHKT6 .6# # 6#6 . .6. 6

NRGGGTGQSATSWDNTATF-KLGVISAWARAT

QGTLLSGWRLTTFMNTVLNWAYMKLAGVFDL-

VII:u:

VTFLKRFFVVFLKRKFVTFLKRHFLTFLKRSFCDFLKRTFLDFLKRFF

IIILWFMSHRALWFMSHRANPYMSLTM6##

VEYLSK--.6YS.AQYLSRSC

RADEKYPFLKKEGP---LHIDYG-TGFNLVED---R-VQRS-STIYVHPEHGQV

LERDG---MLSKDG---IVKTPF---G

---K---K---EG--KREQQ--APIG--IG--VGD-R6GKPP

vImI

IHPVMPMKEIHESIRWYRPVMNREALEAMLSYYKPVMASKTLEAILSFIRPAISEKTIWSLIAWWDAPEDKASLWSQLHYW-APLDKSAIFSCLHW

IYIPKLEEERIVSILEWLIPKLEPERIVSILEWVGFSLPVERIIAIMQW*R S* 6#LPRRPTAEDSADYRAW

ATLVHSRIE SNEPLSWRVME

FIG. 1. Putative RNA-dependent RNA polymerase encoded by the chestnut blight hypovirulence-associated dsRNA. The alignment of theputative HAV polymerase domain with the conserved domains of polymerases from supergroup I positive-strand RNA viruses and theSaccharomyces cerevisiae dsRNA virus ScV is shown. (See abbreviation footnote for names of viruses.) The conserved motifs found in allpositive-strand RNA viral polymerases (I-VIII), as well as a motif specific for the polymerase supergroup I (Ia), are boxed. For the ScVpolymerase, no definite alignment in the region upstream of motif I could be found. For the polymerases of HAV and ScV, the spacers betweensegments IV and V were not aligned; their lengths are indicated. *, Identical amino acid residues in all sequences; :, identical or similar residuesin all sequences; !, identical or similar residues in the sequences of HAV, the potyviruses and BaYMV; *, identical residues in HAV andBaYMV, or in HAV and ScV; ., similar residues in HAV and BaYMV, or in HAV and ScV. The following groups of similar amino acid residueswere considered: G and A; S and T; D, E, N. and Q; K and R; L, I, V, and M; and F, Y, and W. The distances from the protein N-terminito the aligned segments are indicated. Additional sequence, extending past the C-termini of the alignments, is found in each protein.

second highest. Analysis of the surrounding portion of theORF B protein using the SITE program and visual inspectionalso identified the counterparts to six other conserved motifstypical of positive-strand RNA viral RNA polymerases (13-16). Comparisons ofthe HAV polymerase-like sequence withall known sequences of the RNA polymerases of positive-strand RNA and dsRNA viruses showed that the closestsimilarity was indeed with potyviruses and BaYMV, fol-lowed by picorna-, como-, and nepoviruses. Multiple align-ment of these supergroup I (picomavirus-like) positive-strand RNA virus polymerases (16) with the HAV sequencerevealed two segments (motifs Ia-IV and motifs V-VII) thataligned with convincing alignment score (AS) values of 7.2SD and 9.6 SD, respectively. Generally, scores above 7 SDare considered a solid indication of relatedness among se-quences, while scores between 4 and 7 SD are significant,provided that additional evidence to support the suspectedrelationship is available.

In view of the recent suggestion that dsRNA viruses mayconstitute a monophyletic group (17), the possible relation-ship between the putative RNA polymerase ofHAV and theRNA polymerases ofdsRNA viruses was investigated. Initiallocal similarity searches did not reveal specific relationshipsbetween the putative HAV polymerase sequence and any ofthe six known sequences of dsRNA viral polymerases (not

shown). Alignments of the polymerases of dsRNA viruses,supergroup I positive-strand RNA viral polymerases, and theHAV polymerase mostly had AS values of less than 5 SD.These results are exemplified in Fig. 1 by the alignment oftheyeast dsRNA virus (ScV) polymerase. The ScV sequencewas selected among the dsRNA virus polymerases for inclu-sion in Fig. 1 only on the basis that both ScV and HAVreplicate in fungal hosts. It is clear that the overall similaritybetween HAV and ScV was less pronounced than thatbetween HAV, the potyviruses, and BaYMV. Within theregion compared, HAV shared 45 identical and 29 similarresidues with BaYMV (263 residues aligned) versus 24 and 28residues with ScV (259 residues aligned). A unique feature ofthe putative polymerase ofHAV was the insertion of ca. 130amino acid residues between motifs IV and V. Inserts ofsomewhat smaller size have been found previously at similarlocations in the putative polymerases oftwo dsRNA viruses,bluetongue virus and reovirus (13). Another interesting fea-ture of the putative HAV polymerase was the substitution ofserine for glycine in the highly conserved GDD tripeptide(motif VI in Fig. 1). Analogous substitutions occur in theputative polymerases of coronaviruses, toroviruses, severalnegative-strand RNA viruses, and the dsRNA bacteriophage46 (13, 15, 18-20), although no particular sequence similar-ities could be detected between the putative polymerases ofthese viruses and HAV despite an extensive search.

PV1 122EMCV 117FMDV 123HA 118CpMV 144GCMV v

TEV 132TVMV 132BaYMV 159

HAV 1786

ScV 341

PV1EMCVFMDVHACpMVGCMV

TEVTVMVBaYMV

RAV

Dow

nloa

ded

by g

uest

on

Apr

il 29

, 202

0

Evolution: Koonin et al. Proc. Natl. Acad. Sci. USA 88 (1991) 10649

The alignment shown in Fig. 1 was used for the generationof tentative phylogenetic trees using three independent algo-rithms. Additionally, the polymerase of black beetle virus(BBV), demonstrated to be a peripheral member of thepolymerase supergroup I (16), and the polymerases ofdsRNA viruses were included in the analysis. The clusterdendrogram shown in Fig. 2 suggests grouping of the HAVpolymerase with the polymerases of the potyviruses andBaYMV. A similar branching order was obtained when themaximum topological similarity algorithm was used. Theprotein parsimony algorithm produced a somewhat differenttopology, suggesting separation of HAV before the divisionof the potyvirus/BaYMV and picorna/como/nepovirusgroups (not shown). Despite this apparent difference, eachmethod demonstrated the grouping of the HAV polymerasewith the polymerases of positive-strand RNA viruses, andnot with those of dsRNA viruses (Fig. 2).The Putative Helicase of HAV Belongs to the Helicase

Superfamily II and Is Related to its Positive-Strand RNA ViralMembers. Previous analysis detected motifs typical of heli-cases in the C-terminal portion of the ORF B product ofHAV(6) and showed some similarity to the putative RNA helicase(CI protein) of tobacco vein mottling potyvirus (TVMV). Themore detailed analysis performed here confirms that theHAV helicase-like sequence largely conformed to the con-sensus pattern of conserved amino acid residues typical ofthe so-called helicase superfamily II (21), though it containedsome notable deviations (Fig. 3). This superfamily broughttogether the (putative) RNA helicases of poty-, flavi- andpestiviruses, that of hepatitis C virus (22), the "DEAD"family of cellular RNA helicases (with translation initiationfactor eIF-4A as the prototype), and a number of DNAhelicases. We generated the alignment of these helicases withthe putative helicase of HAV, yielding a convincing AS ofover 10 SD (shown in excerpts in Fig. 3). Tentative phylo-genetic trees revealed the grouping of the putative helicase ofHAV with those of positive-strand RNA viruses, though itseemed to be roughly equidistant from the helicases ofpotyviruses and BaYMV, on the one hand, and those of

ROTABTVREOSCV-_IBDVBBVHA V

TVMV

TEVGCMV

CPMVHAPVEMCV

FMDVPhi6

0 1 2

FIG. 2. Cluster dendrogram for the conserved regions of the RNApolymerases of HAV, positive-strand RNA viruses, and dsRNAviruses. The branch lengths are proportional to the distances calcu-lated. A logarithmic scale showing these distances (in relative units)is at the bottom. The branching order in this dendrogram does notnecessarily imply that the polymerases of the majority of dsRNAviruses are monophyletic. Upon further analysis, some may joinother groups of positive-strand RNA viruses that were not analyzedhere.

JEV 187aa QMTVLDLHPGSGKTR-KILPQKUN 187aa QITVLLDLHGAGKTR-RILPQ;,NNV 187aa QITVLDLHPGAGKTR-KILPQYFV l91aa MTTVLDFHPGAGKTR-RFLPQTBEV 191aa QITVLDMHPGSGKTH-RVLPEBVDV 1898aa DFKQITLATGAGKTT-E-LPKHCV 1223aa QVAHLHAPTGSGKST-K-VPATEV 77aa RDFLVRGAVGSGKST-G-LPYPPV 77aa QDILIRGAVGSGKST-G-LPFPVY 78aa LDFLVRGAVGSGKST-G-LPVTVMV 78aa KDIILMGAVGSGKST-G-LPTBaYNV 87aa WSMVV-GHTGSGKST-Y-LPVConsensus ++ G GKS LP

lalOaa RTAVLAPTRVV 50aalOaa RTAVLAPTRVV' 50aa10aa RTAVLAPTRVV 50aalaa RTLVLAPTRVV 50aalOaa RTLVLAPTRVV 50aalOaa RVLVLIPLRAA 54aa7aa KVLVLNPSVAT 49aa6as RVLMLEPTRPL 53aa6aa HVLLIEPTRPT 53aa6aa SVLLIEPTRPL 53aa6aa GVLLLEPTRPL 53aa

14aa QILICEPTQAA 53aa+++ PTR +

THAV 2657aa GHVtVAAKTASGKST-F-fPA 12aa KLWIVMPrKIL 45aaeIF-4A 69aa YDVIAQAQSGTGKTATFAIsI llaa QALVLAPTREL 63aa

II III IVJEV YNLFVMDEAHFT 18aa EAAAIFM-TATPPG 29aa VWFVASVKMGNEIKUN YNLFVMDEAHFT 18aa EAAAIFM-TATPPG 29aa VWFVPSVKMGNEIWNV YNLFIMDEAHFT 18aa EAAAIFM-TATPPG 29aa VWFVPSVKMGNEIYFV WEVIIMDEAHFL 18aa ESATILM-TATPPG 29aa AWFLPSIRAANVMTBEV WEVAIMDEAHWT 18aa KCALVLM-TATPPG 29aa AWFVPSSAKGGIIBVDV YSYIFLDEYHCA 16aa SIRVVAM-TATPAG 32aa LVFVPTRNMAVEVHCV YDIIICDECHST 18aa GARLVVLATATPPG 28aa LIFCHSKKKCDELTEV YDFVIIDECHVN 16aa EGKVLKV-SATPPG 32aa LVYVASYNDVDSLPPV YKCIIFDECHVH 16aa PGKILKV-SATPPG 32aa LVYVASYNEVDTLPVY FNFVIFDECHVL 16aa ACKVLKV-SATPVG 32aa LVYVSSYNEVDTLTVMV YQFIIFDEFHVL 16aa NGKIIKV-SATPPG 32aa LVYVASYNEVDQLBaYMV FDAIFLDEAHDV 15aa SVRKFYV-SATPRD 31aa LVFLAGRPECIKACon Y ++++DE H + + TATP G ++F+ S

S YHAV dNLVFFDEFHEM llaa KGPTIFM-SATPVA 36aa MIiVPTYNELKKTeIF-4A iKMFVLDEAdEM 18aa NTQVVLL-SATmPs 46aa VIFINTRRKVDWL

V VIJEV 18aa VITTDISEMGANF-GAS-RVID 28aa SAAQRRGRVGRNP 153aaKUN 18aa VVTTDISEMGANF-KAS-RVID 28aa SAAQRRGRTGRNP 153aaWNV 18aa VYTTDISEMGANF-KAS-RVID 28aa SAAQRRGRIGRNP 153aaYFV 18aa ILATDIAEMGANL-CVE-RVLD 27aa SAAQRRGRIGRNP 154aaTBEV 18aa VVTTDISEMGANL-DVS-RVID 25aa SAAQRRGRVGRQE 153aaBVDV l9aa IVATNAIESGVTLPDLD-TVID 30aa EQAQRRGRVGRVK 1797aaHCV 17aa VVATDALMTGYTG-DFD-SVID 29aa SRTQRRGRTGRGK 1515aaTEV 22aa IVATNIIENGVTI-DID-VVVD 26aa ERIQKLGRVGRHK 274aaPPV 22aa VVATNIIENGVTL-DID-VVVD 26aa ERIQRLGRVGRNK 275aaPVY 22aa VVATNIIENGVTL-DID-VVVD 26aa ERIQRLGRVGRFK 274aaTVMV 22aa IVATNIIENGVTL-DVD-VVVD 26aa ERIQRFGRVGRNK 275aaBaYMV 24aa IFTTNIIETGVTL-SVD-CVVD 26aa ERQQRIGRAGRLK 282aaCon ++ TDI E G + V+D QR GR+GR

NHAV 17aa LVCTpYVQTGIDIKPAPSILID 22aa TNEQRVnRVGRTM 136aaeIF-4A 22aa LITTDLLaRGIDVQQVS-LVIN 7aa NYIhRIGRgGRFG 39aa

FIG. 3. Conserved motifs typical of the helicase superfamily II inthe putative helicase of HAV. The consensus pattern of conservedamino acid (aa) residues for the positive-strand RNA viral membersof the helicase superfamily II (after ref. 21 with some modifications),the putative HAV helicase, and eukaryotic translation initiationfactor eIF-4A is shown. The HAV-encoded and eIF-4A residues thatdo not conform to the consensus are shown in lowercase. Residuesthat are structurally conserved (see Fig. 1 legend) between theHAV-encoded sequence and most of the other sequences are indi-cated by +.

flaviviruses, pestiviruses, and hepatitis C virus, on the otherhand (not shown). Thus, our analysis demonstrated thepossibility of a common origin for the HAV helicase and theRNA helicases of positive-strand RNA viruses, though itfailed to reveal a close association of the former with aspecific group of the latter. One could argue that the HAVhelicase has diverged far from its positive-strand RNA viralrelatives because of different functional requirements asso-ciated with replication and/or transcription of the HAVdsRNA.HAV Encodes Two Cysteine Proteases and an Additional

Domain Related to the Helper-Component Protein of Potyvi-ruses. Recent studies have revealed that the two HAV-encoded polyproteins undergo autoproteolysis. The productof ORF A is processed into p29 and p40, with the proteolyticactivity residing in p29 (7). A 48-kDa protein, p48, is auto-catalytically cleaved from the N terminus of the ORF Bproduct (6). Site-directed mutagenesis (23) has identifiedCys-162 and His-215 of p29 as residues essential for auto-proteolysis, indicating that this enzyme, like the potyvirus-encoded helper-component (HC) protease (24), resemblespapain-like proteases. Additional similarities between p29and HC protease have been noted previously (7) in the formof conserved amino acid sequences around the essentialcysteine and histidine residues, the nature of the cleavagedipeptides, and the distances between the essential residuesand the cleavage sites. Recent analysis of p48 has also

Dow

nloa

ded

by g

uest

on

Apr

il 29

, 202

0

Proc. Natl. Acad. Sci. USA 88 (1991)

revealed a single cysteine (Cys-341) and a single histidine (His-388) as essential for autoproteolysis (25). A more detailed com-parison of the two proteases of HAV with the potyvirus andBaYMV HC protease domains yielded an alignment includingquite a number of conserved residues, despite its relativelymodest statistical significance, with AS values of 5.1 and 3.9 SDfor p29 and p48, respectively (Fig. 4).The presence of two related proteases among HAV ge-

nome products led Shapira and Nuss (25) to speculate that therespective genes could have evolved by duplication. Tenta-tive phylogenetic trees were generated to assess the related-ness of these proteases to one another and to potyviral HCproteases. The sequences of the C-terminal domains ofalphavirus nsP2 proteins, which are cysteine proteases (26)that display significant similarity to the potyvirus HC prote-ases (A. E. Gorbalenya, E.V.K., and M. M. C. Lai, unpub-lished observations), were included as an outgroup. All threetree-generating algorithms mentioned above agreed in sug-gesting the grouping between the p29 and p48 proteases, asexemplified by the cluster dendrogram in Fig. 5. Thus the twoprotease genes of HAV could have evolved by intragenomicduplication followed by rapid divergence due possibly to thefunctional diversification of the proteases. Despite an exten-sive computer-assisted search, no similarity was detectedbetween any additional HAV-encoded proteins and serine,cysteine, acid, or picornavirus 3C-type proteases.Comparison ofthe sequences of the potyvirus HCs and p29 of

HAV revealed an additional domain of moderate similarity (ASof 5.8 SD) at the N termini of these proteins, marked by theconservation of three Cys residues (Fig. 4). Potyvirus HC pro-teins contain approximately 200 amino acid residues between theN-terminal conserved region and the C-terminal protease do-main, for which there is no counterpart in p29. HC proteins,therefore, may consist of three distinct domains, with the ho-mologs of only the C-terminal protease and the N-terminaldomain present in p29. In contrast, the respective protein ofBaYMV contains only the protease domain and a truncated formof the central domain ofHC (Fig. 4).Comparison of the Genetic Organizations and Expression

Strategies ofHAV, Potyviruses, and BaYMV: A Possible Caseof Evolution by Extensive Genome Rearrangement. The com-parative amino acid sequence analysis described above re-vealed four conserved domains among the gene products ofHAV, potyviruses, and BaYMV (one of which is duplicatedin HAV). Tartaglia et al. (27) speculated previously that

NSP2 SFV

0 1

FIG. 5. Cluster dendrogram for the papain-like proteases ofHAV(48- and 29-kDa proteases), potyviruses, BaYMV, and alphaviruses.NSP2, nsP2 cysteine proteases. For details see legend to Fig. 2.

hypovirulence-associated dsRNAs are analogous to the rep-licative form of an ancestral single-stranded RNA virus.Considering the relative organization of the conserved do-mains within the HAV-encoded, potyvirus-encoded, andBaYMV-encoded polyproteins (Fig. 6), it is tempting tosuggest that HAV-dsRNA could have evolved by rearrange-ment of a positive-strand RNA potyvirus-like genome. Thescenario for this rearrangement could include the followingmajor events (their indicated order is arbitrary): (i) transpo-sition of the helicase gene; (ii) duplication of the sequenceencoding the protease domain of HC; (iii) deletion of thesequence encoding the protease domain of NIa; (iv) deletionof the capsid protein gene; and (v) emergence of the termi-nation codon separating ORFs A and B.For steps iii and iv, the alternative is the loss of the

respective function followed by extensive divergence. Inview of the potential for movement of the ancestral viruswithin the mycelium and the frequency of anastomosis, it isconceivable that an extracellular route of infection, and therequired packaging function, would have been dispensable.Similarly, in the absence of a capsid protein, the ratio of thesingle-stranded RNA genome to the replicative form could havebeen altered so that the double-stranded form predominated.

**:: * *

TEV HC 32Oaa FLELRPDGI-SHECTRGVS-VRRCGZVAAILTQALSPCGGI---TCKRC-MV-KTPDIVEGESGESVTNQGK-LLAML-KEQYPDFPPV HC 322aa FMQCKLRET-DHQSTSDLD-VKECGDVAALVCQAIIPCGRI---TCLQC-AQ-KYSYMSQQEIRDRFSTVIEQHEKTA-MDNYPQFPVY HC 297aa WARMRYPS--DHTCVAGLP-VEDCGRVAALIAHSILPCYKI---TCPTC-AQ-QYASLPVSDLFKLLHKHARDGLNRL-GADKDRFTVMV HC 269aa FQEQKAIGL-DHTCTSDLP-VEACCHVAALJCQSLFPCGKI---TCKRC-IA-NLSNLDFDTFSELQGDRAMRILDVM-RARFPSFPSBMV HC 41laa YVKLRDESVSDHDCVGGIT-PEECGILAAQILRVFYPCWRI---TCTKC-IS-NWLSKPTSEQIEHIYRRGNLAIQDL-NKRIPSAHAV P29 24aa FVSVRTKEVVPAGCITLWEYRDSCGDVPGPLSRQDLRRLRTPDGVC-KCQVHFKLPTVLKSGSTGTVPKHPAVLAAFIGRPRRCSL

PMAEKLLT-RFLQQKSLV 199aaSHVLAFLK-RYRCLMRVE 199aaIHVNKFLI-ALEHLETPV 200aaTHTIRFLH-DLFTQRRVT 199aaHHVTQMVE-LLRQRIKNT 197aa

107aaEQRTKELDSRFLQLVHGG Oaa

306aa

HPTKRHLVIQNSGDSKYLD--LPVLNIEKKYIANIGYCXINIFFALLVNVKKEDAKDFTKFIRDTISPTKNHLVVGNTGDSKYVD--LPTAKGGAMFIAKAGYCYINIFLAHLININEDEAKSFTKTVRDTLPPTKKHLVIGNSGDQKFVD--LPKGDSEHLYIAKQGYCYINVFLANLINISEEDAKDFTKKVRDMCMPTKNHLVIGNSGDPKYLD--LPGIISNIMYIAKIGYCYINIFLANLVNVDEANAKDFTKRVRDESLPTRNHLVIGNTGDPKLVD--LPKTETGRBWIAKEGYCYINIXFAXLVNVSZKDAKDFTKFVRDEISHAPWLYMA-NNVCAYEAT-HLKPVQTFIAFDFAHQYCYLSLFIPLSFRITPENARSFSRFL-EQLLPARPSYMIARPPRP-VRG--LCSSRNGSLAQFGQQYCYLS---AIVDSARWRVARTTGWCVRVADAP-RQQLVPRRST---FVDNHEEVKIDTLRVPVIEGRCF-----ELLFN-NQVTPAIFDK--KPLL

*

* : : *: * :VPKLGAWPTMQDVATACYLLSILYPDVLRAZLPRILVDHDNKTMHVLDSYGSRTTGYHMLKMNTTSQLIEFVHSGLKSEMKTYNVGVPKIGTWPSMMDLATACHFLAVLYPETRNAELPRILVDHEAKIFHVVDSFGSLSTGMHVLKANTINQLISFASDTLDSNMKTYLVGVPKLGTWPTNMDLATTCAQMRIFYPDVHDAELPRILVDHDTQTCHVVDSFGSQTTGYHILKASSVSQLILFANDELKSDIKHYRVGVQKLGKWPSLIDVATECALLSTYYPAAASAELPRLLVDRAQKTINVVDSYGSLNTGYHILKANTVSQLEKFASNTLESPMAQYKVGMPQLGKWPTMMDVATACYKLAIIYPDVRDAQLPRILVDRSEQIFWVIDSYGSKTTGYHILKAGTVSQLISFXHGALLGEMKMYRVGPDILGAYPTLAAIYKTMLFAIRLFPEVLQAPIPIIAKRPGVLQFRVSDARG-LPPSWFPMWCGSVRSFVALITNNLNSDLLDGIVGYLRLLQWVGRRSFGS--------F------QIEKSAVDH---VYBW-----VDAZYQSEQDGALFYQAIL---GLAEKDPLARIGKDVLGVFEEV------C-------------TMDSLEISHSDQCVHIVA--GETFRNYDEIK--AVLEVI------LKNE-PDILVG

FIG. 4. Two papain-like proteases and the HC-related domain of HAV. The alignments with two domains of potyvirus-encoded HC proteinsare shown. The distances between the two HC domains, and the distances between the N terminus of each polyprotein and the indicatedsequence, are given. The C-terminal boundaries of the aligned segments correspond to the cleavage sites. The N-terminal sequences of theHC-related protein of BaYMV and of p48 are not shown, as they displayed no detectable similarity to any portion of the potyvirus proteins orto each other. *, Identical residues in all sequences; :, similar residues in all sequences; *, the (putative) catalytic residues of the proteases;boldface, individual residues that are similar or identical to HAV residues.

TEV HCPVY HCPPV HCTVMV HCPSBMV HCBaYMV'HC'RAV P29HAV P48

TEV HCPVY HCPPV HCTVMV HCPSBMV HCBaYMV'HC'HAV P29HAV P48

10650 Evolution: Koonin et al.

Dow

nloa

ded

by g

uest

on

Apr

il 29

, 202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 10651

k1 '(: 1:( \ 1 1 !'

N -,ii

Ii.LI7HiieiIi7_ ______- §71L D D kII1AN'~~ ~ ~ ~ I 111i i

LaidL__ 7]- __eA-Ja *_-iIii1L,1 OK \V P-,-P

R\ \ PI'i i \R N. Pk r,

FIG. 6. Schematic comparison of the organization of the poly-proteins of HAV, potyviruses, and BaYMV. Related domains are

highlighted by identical shading. The termination codon of the HAVORF A is indicated by a solid vertical line, and the cleavage sites are

denoted by dashed lines. Pol, RNA polymerase; Hel, helicase; Pro,protease; CP, capsid protein. The BaYMV scheme was designed bycombining data from refs. 8, 9, and 12.

Step v is parallel to the possible evolution of BaYMV, whichappears to have involved the isolation of the 5'-terminal genes ina separate RNA segment. It should be noted that BaYMV istransmitted in nature by a fungal vector (28), although the abilityofBaYMV to replicate in the vector has not been demonstrated.The order of the helicase and polymerase domains observed inthe potyviruses and BaYMV (N-. . . heicase-. . . -polymerase-... -C) is typical of a wide range of positive-strand RNA viruses(21, 29). Interestingly, the reversed order observed in HAV hasalso been found in corona- and toroviruses (18-20). Moreover,the overall organization of the HAV L-dsRNA is surprisinglysimilar to that of coronaviruses, with coding sequences for twopapain-like proteases in the 5'-proximal part of the RNA andsequences encoding the polymerase-helicase array in the 3'-proximal region (30). However, upon careful comparison, we

have not found any specific sequence similarities between pro-

teins encoded by HAV and the coronaviruses. Thus we believethat the striking analogy in the genetic organizations of theseviruses is a peculiar case of evolutionary convergence.

IfHAV and potyviruses had a common origin, was the hostof the ancestral virus a plant or a fungus? Perhaps the mostplausible series of events involves the acquisition of a poty-virus-like virus by the fungus during saprophytic or patho-genic interactions with a plant host. Since the HAV dsRNAconfers a hypovirulent phenotype to C. parasitica, acquisi-tion of an ancestral virus may have provided a survivaladvantage to both the plant host and fungal pathogen. Alter-natively, a fungal host could have acted as a vector, trans-mitting an ancestral dsRNA virus to a plant host followed bysubsequent evolution to a single-stranded RNA virus as

suggested by Bruenn (17). These considerations are compli-cated because of the apparent lack of an encapsidated fromof the HAV RNA, prohibiting a clear distinction betweengenomic and nongenomic RNA in the classical sense. Nev-ertheless, in considering the evolutionary relationship be-tween single-stranded and double-stranded RNA viruses, it isinteresting to note that HAV dsRNA frequently undergoesinternal deletion events and that recombination with cellularRNA may also occur (31). This propensity for recombinationis compatible with the proposed genome reorganization in thecourse of evolution of HAV dsRNA from a potyvirus-likeancestor. Moreover, one can envision that additional internaldeletion events could lead to the formation of segmenteddsRNA genomes from HAV-like genetic elements. Previ-ously, gene module shuffling has been highlighted as a majortrend in the evolution ofpositive-strand RNA viruses (32, 33).Here, we demonstrate that a similar process may account forthe evolution of a dsRNA virus-like genetic element from a

positive-strand RNA virus.

E.V.K. is most grateful to Dr. A. E. Gorbalenya for helpful discus-sions and for permission to use his unpublished observations. Thanks are

also due to Dr. V. V. Doija for critical reading of the manuscript and toDrs. A. Davidson and S. Kashiwazaki for providing preprints of theirwork. The excellent assistance of Ms. Ula Dodson and Ms. AshleyHenderson in the preparation of this manuscript is appreciated. Thiswork was supported in part by grants from the National Institutes ofHealth (AI27832), the U.S. Department of Agriculture (89-37263-4503),and the Texas Advanced Technology Program, and by funds provided bythe Department of Biology, College of Science, and Provost's Office ofTexas A&M University.

1. Anagnostakis, S. L. (1982) Science 215, 466-471.2. Van Alfen, N. K. (1982) Annu. Rev. Phytopathol. 20, 349-362.3. Roane, M. K., Griffin, G. J. & Elkins, J. (1986) Am. Phyto-

pathol. Soc. Monogr., pp. 1-53.4. Griffin, G. J. (1986) Hort. Rev. 8, 291-335.5. Nuss, D. L. & Koltin, Y. (1990) Annu. Rev. Phytopathol. 28,

37-58.6. Shapira, R., Choi, G. H. & Nuss, D. L. (1991) EMBO J. 10,

731-739.7. Choi, G. H., Shapira, R. & Nuss, D. L. (1991) Proc. Natl.

Acad. Sci. USA 88, 1167-1171.8. Kashiwazaki, S., Minobe, Y., Omura, T. & Hibino, H. (1990)

J. Gen. Virol. 71, 2781-2790.9. Kashiwazaki, S., Minobe, Y. & Hibino, H. (1991) J. Gen. Virol.

72, 995-999.10. Johansen, E. (1990) PhD thesis (Institute for Ecology and

Molecular Biology, Copenhagen).11. Dolja, V. V., Boyko, V. P., Agranovsky, A. A. & Koonin,

E. V. (1991) Virology 184, 79-86.12. Davidson, A. D., Prols, M., Schell, J. & Steinbiss, H.-H. (1991)

J. Gen. Virol. 72, 989-993.13. Koonin, E. V., Chumakov, K. M. & Gorbalenya, A. E. (1989)

FEBS Lett. 252, 42-46.14. Koonin, E. V., Gorbalenya, A. E., Chumakov, K. M.,

Donchenko, A. P. & Blinov, V. M. (1987) Mol. Genet. Mikro-biol. Virusol. 7, 27-39.

15. Pock, O., Sauvageut, I., Delarue, M. & Tordo, N. (1989)EMBO J. 8, 3867-3874.

16. Koonin, E. V. (1991) J. Gen. Virol. 72, 2197-2206.17. Bruenn, J. A. (1991) Nucleic Acids Res. 19, 217-226.18. Gorbalenya, A. E., Koonin, E. V., Donchenko, A. P. & Bli-

nov, V. M. (1989) Nucleic Acids Res. 17, 4456-4469.19. Bredenbeek, P. J., Pachuk, C. J., Noten, A. F. H., Charite, J.,

Luytjes, W., Weiss, S. R. & Spaan, W. J. M. (1990) NucleicAcids Res. 18, 1825-1832.

20. Snijder, E. J., Den Boon, J. A., Brendenbeek, P. J., Horzinek,M. C., Rijnbrand, R. & Spaan, W. J. M. (1990) Nucleic AcidsRes. 18, 4535-4542.

21. Gorbalenya, A. E., Koonin, E. v., Donchenko, A. P. & Bli-nov, V. M. (1989) Nucleic Acids Res. 17, 4713-4730.

22. Miller, R. M. & Purcell, R. M. (1990) Proc. Natl. Acad. Sci.USA 87, 2057-2061.

23. Choi, G. H., Pawlyk, D. M. & Nuss, D. L. (1991) Virology 183,747-752.

24. Oh, C.-S. & Carrington, J. C. (1989) Virology 171, 692-699.25. Shapira, R. & Nuss, D. L. (1991) J. Biol. Chem. 266, 19419-

19425.26. Hardy, W. R. & Strauss, J. H. (1989) J. Virol. 63, 4653-4664.27. Tartaglia, J., Paul, C. P., Fulbright, D. W. & Nuss, D. (1986)

Proc. NatI. Acad. Sci. USA 83, 9109-9113.28. Mathews, R. E. F. (1991) Plant Virology (Academic, San Di-

ego), 3rd Ed.29. Gorbalenya, A. E., Blinov, V. M., Donchenko, A. P. & Koo-

nin, E. V. (1989) J. Mol. Evol. 28, 256-268.30. Lee, C.-J., Shieh, C.-K., Gorbalenya, A. E., Koonin, E. V.,

La Monica, N., Tuler, J., Bagdzhadzhyan, A. & Lai, M. M. C.(1991) Virology 180, 567-582.

31. Shapira, R., Choi, G. H., Hillman, B. I. & Nuss, D. L. (1991)EMBO J. 10, 741-746.

32. Zimmern, D. (1988) in RNA Genetics, eds. Holland, J. ..,

Domingo, E. R. & Ahlquist, P. (CRC, Boca Raton, FL), pp.

211-240.33. Morozov, S. Yu., Dolja, V. V. & Atabekov, J. G. (1989) J.

Mol. Evol. 29, 52-62.

Vo12N iru s

I- (C o .....-;

Evolution: Koonin et al.

Dow

nloa

ded

by g

uest

on

Apr

il 29

, 202

0